skills/0xkynz/codekit/zvec/Gen Agent Trust Hub

zvec

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The documentation provides instructions to install the zvec library from official package registries (PyPI and NPM) and references source code hosted in the Alibaba organization's GitHub repository.
  • [COMMAND_EXECUTION]: Includes examples for database lifecycle management, such as creating collection directories on the local filesystem and performing irreversible deletions of data folders via the destroy() method.
  • [REMOTE_CODE_EXECUTION]: Mentions embedding and reranking functions that download pre-trained model weights from established community registries like HuggingFace or ModelScope upon initial execution.
  • [CREDENTIALS_UNSAFE]: Security best practices are followed by using descriptive placeholders (e.g., 'your-openai-api-key') in configuration examples instead of hardcoded credentials.
  • [SAFE]: The skill facilitates the use of a legitimate vector database tool, with all external references targeting well-known technology organizations and official service endpoints.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 09:08 AM